1
Journal of Exercise Physiologyonline
Volume 14 Number 1 February 2011
Editor-in-Chief
Tommy Boone, PhD, MBA
Review Board
Todd Astorino, PhD
Julien Baker, PhD
Steve Brock, PhD
Lance Dalleck, PhD
Eric Goulet, PhD
Robert Gotshall, PhD
Alexander Hutchison, PhD
M. Knight-Maloney, PhD
Len Kravitz, PhD
James Laskin, PhD
Yit Aun Lim, PhD
Lonnie Lowery, PhD
Derek Marks, PhD
Cristine Mermier, PhD
Robert Robergs, PhD
Chantal Vella, PhD
Dale Wagner, PhD
Frank Wyatt, PhD
Ben Zhou, PhD
Official Research Journal
of the American Society of
Exercise Physiologists
ISSN 1097-9751
JEPonline
Resistance Exercise Energy Expenditure is Greater with
Fatigue as Compared to Non-Fatigue
CHRISTOPHER B. SCOTT1, CONRAD P. EARNEST2
1
University of Southern Maine, Gorham, USA, 2Pennington Biomedical
Research Center, Baton Rouge, USA
ABSTRACT
Scott CB, Earnest CP. Resistance Exercise Energy Expenditure is
Greater with Fatigue as Compared to Non-Fatigue. JEPonline
2011;14(1):1-10. We retrospectively investigated data from two separate
studies to estimate and compare aerobic and anaerobic exercise energy
expenditure (EE) along with the aerobic recovery EE component for 1set of resistance exercise. One study was completed using non-fatiguing
lifts where the exercise was stopped before muscular failure. In another
study muscular failure (fatigue) was the end point of all lifts. Work
(weight lifted × upward vertical displacement) and all EE components
were examined. Non-fatiguing lifts were carried out at 50% of a 1-RM for
7, 14 and 21 repetitions. Lifts to failure were carried out at ~37%, ~46%,
~56%, 70%, 80% and 90% of a 1-RM. Individual regression lines were
created for fatigue and non-fatigue conditions for each male subject
between work and all estimates of EE. The results of our analyses
showed that the averaged slopes between fatigue and non-fatigue were
proportional for: total EE/work (p = 0.87), anaerobic exercise EE/work, (p
= 0.73) and recovery EE/work (p = 0.19). However, the Y-intercepts of
the two studies were significantly greater for fatiguing as compared to
non-fatiguing lifting for: total EE/work (p = 0.007), anaerobic exercise
EE/work (p = 0.001) and recovery EE/work (p = 0.01), but not aerobic
exercise EE/work (p = 0.17). For aerobic exercise EE/work, lifting to
fatigue had a greater O2 uptake/work slope as compared to lifts that
were not completed to fatigue (p = 0.04). We conclude that lifting a
weight to muscular failure can entail significantly greater aerobic,
anaerobic and recovery EE components as compared to non-fatiguing
lifting.
Keywords: Anaerobic Energy Expenditure, Oxygen Uptake, EPOC,
Lactate, Resistance Training
2
INTRODUCTION
Resistance training is considered an intense form of exercise when sets are performed to or close to
muscle failure, regardless of the percentage of a one repetition maximum (1-RM) selected. To the
contrary, when a completed set of lifted weights does not approach fatigue the exercise can be
considered light-to-moderately heavy. Though this may be intuitively obvious, the distinction between
each form of exercise may be critical for exercise programming depending on the goals of the
participants. For example, recent studies have shown that self-selected resistance training intensities
are lower then what is recommended (3,4,8) and a better understanding of the energetic demands
associated with various resistance training intensities will ultimately help resistance training
programming efforts.
Two contemporary studies using disparate energy expenditure methodologies – steady state and
non-steady state - each report energy expenditure relationships with resistance training-type work
that appear much higher than past estimates (9,10). We reason that the energy expenditure
characteristics of steady rate aerobic-type exercise should not be used to model the energy
expenditure of non-steady rate anaerobic-type exercise because the physiological and metabolic
characteristics for each are different. When lifting a moderately-heavy to heavy weight for example,
we rationalize that: 1) intense muscle contractions create enough force to effectively limit blood flow
to and from working skeletal muscle, impeding exercise O2 uptake (14), 2) anaerobic glycolytic ATP
production can account for a significant amount of overall energy expenditure (10-13), 3) a single
bout of resistance training results in a greater O2 uptake in the recovery from exercise than during the
actual exercise itself (10,11), and 4) heavy to severe exercise can induce “extra” increases in aerobic
and anaerobic energy expenditure, likely related to changing metabolism-work efficiency and/or the
increased recruitment of muscle needed to prolong fatigue (1,15).
The primary aim of this retrospective investigation was to compare two previous studies that utilized
non-steady state methods of energy expenditure estimation: one where a single set of lifting was
completed before muscle fatigue took place (10), the other where muscular failure was the end point
of the set (11). We asked the question: Are the energy expenditure-to-work relationships different
between single sets of fatiguing and non-fatiguing resistance training protocols?
METHODS
Data collected from two previous studies were examined (10,11). Both studies used 1-set of bench
press exercise to determine the energy expenditure characteristics accompanying non-muscular and
total muscular fatigue (these were not training studies). Comparisons were made of the aerobic,
anaerobic and recovery energy expenditures of each study.
Subjects
Each protocol had received previous approval by the University of Southern Maine’s Institutional
Review Board (IRB). In addition, the current retrospective study underwent further review and
approval by the IRB. For the present investigation only data from male subjects were compared as
men and women can have different aerobic and anaerobic responses to resistance training (see
Table 1).
3
Table 1. Subject characteristics: all males (mean ± SD)
N data pts Age(yrs)
Height (cm)
Weight (kg)
1-RM (kg)
Fatigue study
13
78
23.8 ± 2.1
178.7 ± 6.6
85.9 ± 11.3
102.5 ± 20.8
Non-Fatigue
study
4
36
32.8 ± 10.4
177.5 ± 9.5
76.8 ± 13.1
108.0 ± 15.0
Table 1 legend: N = subject number; data pts = number of lifts measured for each study
Procedures
In the non-fatigue study (10), exercise was performed at 50% of 1-RM for 7, 14 and 21 repetitions
(exhaustion was not the end point of any of these lifts); each workload was performed 3 times for a
total of 9 lifts (36 total data points). In the fatigue study (11), a single set of the bench press was
performed until muscular failure at the following percentage of 1-RM (repetition number at
exhaustion), ~37% (~37 reps), ~46% (~26 reps), ~56% (~20 reps), 70% (~12 reps), 80% (~8 reps)
and 90% (~5 reps); each lift was performed once (78 total data points). Work was recorded as the
product of force (kilograms of weight lifted) × vertical (upward) displacement of lifting a bar using a
standard Smith Machine; the distance the bar travelled was recorded electronically. To determine
energy expenditure for both studies we performed separate measurements of aerobic exercise O2
uptake, blood lactate concentrations and a modified excess post-exercise oxygen consumption
(EPOC) measurement; the summation of all three measures provide an estimate of total energy
expenditure.
Oxygen uptake was measured via a Parvomedics MMS-2400 metabolic cart (Sandy, UT) in 15second sampling intervals. Before lifting a 5-minute supine resting energy expenditure (REE)
measurement was taken with the subject lying on the bench with both feet on the ground. The
average O2 uptake (l min-1) of this rest period was subtracted from all exercise O2 uptake and EPOC
measurements. Exercise O2 uptake underwent conversion as 1 liter of O2 = 21.1 kJ. EPOC was
recorded with the subjects feet elevated horizontal to the bench so that the subject was completely
supine. EPOC was measured until falling below 5.0 ml·kg-1·min-1 (a typical standing resting measure)
or below the averaged measured REE and underwent conversion as 1 liter of O2 = 19.6 kJ. Blood
lactates were collect via a minimum of two Lactate Pro analyzers (FaCT Canada Consulting) at rest
and from a peak lactate measurement taken at 2-minutes or 4-minutes post-exercise (whichever was
highest). Anaerobic exercise energy expenditure was calculated as the difference between resting
and peak lactate values multiplied by body weight (kg), then by 3.0 ml of O2 (2). This O2 equivalent
estimate was converted to Joules as 1 L O2 = 21.1 kJ. Total energy expenditure was recorded as the
sum of aerobic and anaerobic exercise energy expenditures and EPOC.
Statistical Analysis
Energy expenditure measures and work regression lines were computed for each subject, and then
averaged with the slope and Y-intercept data presented in Table 2. Comparisons were made using a
standard t-test (with alpha level set at p < 0.05).
4
Table 2. Averaged individual slopes and Y-intercepts of fatigue (F) and non-fatigue (NF)
studies (mean ± SD)
Study
Anaerobic
Exercise O2
EPOC
Total EE
/work
/work
/work
/work
Slope F
0.049 ± 0.01
0.029 ± 0.01
0.018 ± 0.02
0.084 ± 0.03
Slope NF
0.046 ± 0.02
0.017 ± 0.004
0.024 ± 0.004
0.086 ± 0.02
P value
0.73
0.04
0.19
0.87
Y-int F
8.12 ± 3.9
-3.27 ± 3.9
17.4 ± 7.7
21.2 ± 10.4
Y-int NF
-1.4 ± 2.9
-0.31 ± 1.9
6.08 ± 1.8
4.4 ± 2.8
P value
0.007
0.001
0.17
0.01
Table 2 legend. Slope and Y-intercept represent individual subject relationships between energy expenditure
and work that were subsequently averaged (F, n = 13 men, 6 workloads each performed once, reference 11;
NF, n = 4 men, 3 workloads each performed 3 times, reference 10); TEE = total energy expenditure; work =
force × vertical distance; Anaerobic = anaerobic exercise energy expenditure; Exercise O2 = aerobic exercise
energy expenditure; EPOC = excess post-exercise oxygen consumption energy expenditure; significant
differences between F and NF are shown in bold print.
RESULTS
Data reflecting differences in energy expenditure characteristics are presented in Table 2. Regarding
the amount of work performed, the slopes of the fatigue and non-fatigue regression lines were not
significantly different for total energy expenditure (p = 0.87; Figure 2), anaerobic exercise energy
expenditure (p = 0.73) and recovery energy expenditure (p = 0.19). The two slopes were different for
aerobic exercise energy expenditure (p = 0.04), being greater for the fatigue study. The Y-intercept
data were significantly greater for the fatigue as compared to non-fatigue study for total energy
expenditure (p = 0.007; Figure 2), anaerobic exercise energy expenditure (p = 0.001) and recovery
energy expenditure (p = 0.01); the Y-intercepts were not different for aerobic exercise energy
expenditure.
DISCUSSION
Our intent was to examine the energy costs including the aerobic and anaerobic contributions of lifting
a weight, along with the recovery from that lift. These were not training studies. Indeed, rather than
examine a complete workout, we choose to quantifying resistance training energy expenditure one
exercise at a time (future studies will add sets and exercises to investigate how this subsequently
effects aerobic, anaerobic and recovery energy costs). Our approach allows us to witness unique
differences between steady state and non-steady state exercise. As an example, with 1-set of nonsteady state resistance exercise, exercise O2 uptake always represents the lowest measured energy
expenditure as compared to anaerobic and EPOC energy expenditures (10,11). To the contrary
5
under steady state conditions energy expenditure is defined in terms of exercise O2 uptake. Also,
after a single set of lifting to fatigue, EPOC energy expenditure appears as a relatively constant
quantity unrelated to the volume of work completed (11); with aerobic exercise, EPOC “size” is
related to duration and intensity (6,7).
For a single set of resistance training comparing a series of repetitions carried out to fatigue with a
series of repetitions not performed to fatigue, our data indicate total energy expenditure, including
anaerobic energy expenditure and recovery energy expenditure, are larger for fatiguing as compared
to non-fatiguing exercise as indicated by the Y-intercept data (Table 2). It is further apparent that the
ratio of aerobic energy expenditure to work significantly increased with exercise to fatigue (Table 2).
From these data we suggest that the energy expenditure of fatiguing resistance exercise must be
modeled differently as compared to non-fatiguing resistance exercise.
Exercise O2 uptake
The gold standard modeling of exercise energy expenditure with work applies specifically too easy to
moderately intense aerobic exercise when steady rate power outputs, steady state O2 supply and
steady state O2 uptakes are achieved. Because muscular endurance and strength resistance
exercise do not fulfill these criteria, we describe non-steady state energy expenditure for a given work
bout [in kJoules (kJ)] and not steady state energy expenditure over a per minute time period of power
output (kJ min-1).
As compared to lighter workloads, heavy to severe steady state aerobic exercise reveals an “extra”
energy expenditure component - a slow O 2 component - that is thought to be caused by a decrease in
efficiency and/or the increased recruitment of muscle fibers in an attempt to resist fatigue and prolong
work (1,15). The increasing aerobic exercise energy expenditure - work relationship (slope) for
fatiguing as compared to non-fatiguing lifting may likewise, but not necessarily, be related to
increased muscle recruitment as non-steady rate work is prolonged (lasting minutes). At the Yintercept, a significant change in aerobic energy expenditure was not found between fatiguing and
non-fatiguing exercise (each lasting seconds). As lifting continues and fatigue approaches those
muscles involved with body positioning and placement would likely contribute to the disproportionate
rise in exercise O2 uptake with work output, more intense contractions by the muscles doing the
actual lifting would have impeded blood flow (14). The non-proportional rise in O2 uptake with
fatiguing as compared to non-fatiguing resistance exercise indicates that the aerobic energy
expenditure of the exercise is not consistent as repetitions continue.
Anaerobic Energy Expenditure
Our method of determining anaerobic energy expenditure is based on the difference between peak
and resting blood lactate measurements in the estimation of (glycolytic) anaerobic exercise energy
expenditure (2). As with non-steady state exercise O2 uptake, our anaerobic data provides
information regarding a given work bout (kJ) and not work rate or power output over time (kJ min-1).
Correlation between anaerobic energy expenditure and work was good for both fatigue (r = 0.79, p <
0.0001) and non-fatigue (r = 0.95, p < 0.05) studies. The slope data indicate that the increase
(change) in anaerobic energy expenditure with work when lifting to muscular failure is proportional to
non-fatiguing lifts. However, the greater Y-intercept data for fatiguing as compared to non-fatiguing
lifts reveals a significantly larger anaerobic contribution to energy expenditure for the single set lifts to
fatigue. It appears that a single set of lifting to fatigue results in an “extra” anaerobic energy
expenditure component as compared to non-fatiguing conditions.
6
Excess Post-exercise Oxygen Uptake (EPOC)
After a single bout of resistance training EPOC exceeds exercise O2 uptake, whether fatigue is
involved or not (10,11). In this regard both the exercise and recovery energy expenditures are
required to better portray the energy demands associated with lifting (our O2 uptake to energy
expenditure conversion for EPOC dismisses anaerobic glycolytic ATP re-synthesis; 13,14).
Figure 1.
Figure 1 legend. Excess post-exercise oxygen consumption data is shown for the fatigue study (11). Note that
while the regression line indicates EPOC energy expenditure rises with work and the correlation is statistically
significant (p < 0.05), it also is a poor correlation (r = 0.35). ANOVA analysis indicated no significant
differences in the EPOC energy expenditure among lifts to fatigue at percentages of a 1-RM that ranged from
37% to 90% (nor were differences detected in EPOC among non-fatigue lifts at 50% of 1-RM for 7, 14 and 21
reps; however, EPOC did differ between fatigue and non-fatigue protocols; see ref. 10).
It has been mentioned that the larger the metabolic disturbance of exercise the greater the EPOC will
be, but this consideration is often reserved for aerobic-type exercise where exercise duration (time)
and intensity (% of VO 2max) have a significant impact on EPOC (7). Anaerobic-type exercise also
has duration (time) and intensity (% of 1-RM) components. Yet %VO 2 max and %1-RM do not appear
to be compatible descriptions of intensity in relationship to EPOC. Indeed, aerobic exercise duration
increases EPOC in a linear fashion while aerobic exercise intensity is thought to exponentially
increase EPOC volume (6). With fatiguing anaerobic-type exercise, ANOVA testing indicated that
EPOC’s were in fact similar among all lifts to fatigue (lasting seconds to minutes) as no statistical
differences were seen among lifting intensities that ranged from ~37 to 90% of 1-RM consisting of 4
to 37 reps (11). For the non-fatigue lifts ANOVA again indicated no difference among EPOC’s (50%
of 1-RM lifting consisting of 7, 14 and 21 reps (10). Haddock and Wilkin likewise saw no increases in
7
EPOC-related energy expenditure between 1-set of lifting and 3-sets of lifting to fatigue even though
the volume of work was tripled for the latter (5).
EPOC data for both studies was purposely collected in the short term, until a standing resting O2
uptake was achieved. The lifting to fatigue data reveal poor but significant correlation between EPOC
energy expenditure and weight lifting work (r = 0.35, p = 0.002; Figure 1), as well as between EPOC
and aerobic (r = 0.27, p = 0.02) but not EPOC and anaerobic (r = 0.22, p = 0.06) exercise energy
expenditures. The non-fatigue lifting study on the other hand reveals good correlation between EPOC
and anaerobic energy expenditure (r = 0.75, p < 0.0001) along with EPOC and work (r = 0.83, p <
0.0001), but not with EPOC and aerobic exercise energy expenditure (r = 0.17, p = 0.42). Meirelles
and Gomes (6) indicate a paucity of data and subsequent knowledge regarding EPOC and resistance
training related energy expenditure, we concur. Within groups the question arises, “which is more
important for interpretation purposes: ANOVA comparisons that reveal statistical similarity and
therefore no differences among EPOC’s after various lifting bouts or, statistically significant
correlation between EPOC and lifting that indicate an increase in EPOC with work?” While we cannot
Figure 2.
Figure 2 legend. Total energy expenditure and work. All data are for 1-set of the bench press exercise. The
data for the bottom line (open circles) were obtained from lifts that did not result in fatigue. The top line (closed
circles) was obtained from lifts completed to fatigue. The slopes of the two lines above are not significantly
different (p = 0.87). However, the Y-intercepts of each line are significantly different (p = 0.007). We conclude
that non-fatiguing lifts cannot be utilized to estimate the total energy expenditure (TEE) of lifting to fatigue (TEE
= anaerobic and aerobic exercise energy expenditure + recovery energy expenditure). Non-fatigue, TEE = 4.4
+ (0.086 × work); fatigue, TEE = 21.2 + (0.084 × work).
8
answer this question specifically, between group analyses of the Y-intercept between fatigue and
non-fatigue studies lead us to the conclusion that the short-term recovery (EPOC) energy expenditure
contributions after a single set of fatiguing and non-fatiguing resistance training are different.
Total energy Expenditure
The total energy expenditure of non-steady state lifting consists of aerobic and anaerobic exercise
energy expenditure components along with a modified aerobic recovery energy expenditure (EPOC).
Figure 2 reveals that a composite of the three energy expenditure components is not only different
between 1-set of lifting among fatigue and non-fatigue protocols, but also that the difference is
somewhat proportional throughout a wide range of lifting intensities and work (respectively, different
Y-intercept, similar slope; Table 2). In fact, at a work output of 100 J, there is an estimated 56%
difference in total energy expenditure between non-fatigue (13 kJ) and fatigue lifts (29.6 kJ); at a work
output of 500 J, non-fatigue total energy expenditure (47.4 kJ) is 25% less than lifting to fatigue (63.7
kJ). We conclude that with equivalent single sets of resistance training work, fatiguing exercise results
in greater total energy expenditure as compared to non-fatiguing exercise.
The Y-intercept data of EPOC energy expenditure and work was significantly greater but the slope of
EPOC and work was similar for fatigue and non-fatigue weight lifting protocols. The same was true for
anaerobic energy expenditure and work as well as total energy expenditure and work. Why is this?
We speculate that the anaerobic, aerobic recovery (EPOC) and total energy expenditure’s for 1-set of
resistance exercise to fatigue may be related to a capacity or limit of muscle recruitment along with an
equivalent utilization of ATP, creatine phosphate (CP) and oxygen stores, regardless of work output
or the aerobic and anaerobic (glycolytic) energy expenditure contributions that differ, often
dramatically, both within and between lifting protocols. As an example, because anaerobic energy
expenditure (based on blood lactate) and EPOC are not well related, 37 reps at 37% of a 1-RM at
fatigue recruits a similar amount of muscle mass and uses a similar amount of stored ATP, CP, and
oxygen as does 8 reps at 80% of 1-RM at fatigue, resulting in a similar short-term EPOC. It is of
further interest to determine if or how additional sets and exercises affect EPOC.
CONCLUSIONS
These retrospective data reveal that with 1-set of non-steady state resistance exercise, O2 uptake
increases disproportionately with fatiguing as compared to non-fatiguing exercise. Moreover,
anaerobic, aerobic recovery (EPOC), and total energy expenditures are significantly greater with
resistance work involving fatigue as compared to non-fatigue. It is concluded that energy expenditure
modeling for fatiguing resistance exercise cannot be based on non-fatiguing measurements. For
anaerobic-type exercise to fatigue we continue to propose that a valid estimate of energy expenditure
cannot be based on steady-state O2 uptake measurements and must include the sum of aerobic and
anaerobic exercise and aerobic recovery energy expenditures.
9
Address for correspondence: Scott CB, PhD. Department of Exercise, Health and Sport Sciences,
University of Southern Maine, Gorham, ME, USA, 04038. Phone (207)780-4566 FAX: (207)780-4745;
Email. [email protected].
REFERENCES
1. Aagaard P, Bangsbo J. The muscular system: design, function, and performance relationships. In:
ACSM’s Advanced Exercise Physiology Philadelphia: Lippincott Williams and Wilkins, 2006.
144-160.
2. di Prampero PE, Ferretti G. The energetic of anaerobic muscle metabolism: a reappraisal of older
and recent concepts. Resp Physiol 1999:118;103-115.
3. Focht BC. Perceived exertion and training load during self selected and imposed-intensity
resistance exercise in untrained women. J Strength Cond Res 2007; 21:183-187.
4. Glass S, Stanton D. Self-selected resistance training intensity in novice weightlifters. J Strength
Cond Res 2004;18:324-327.
5. Haddock BL, Wilkin LD. Resistance training volume and post exercise energy expenditure. Inter J
Sports Med 2006;27:143-148.
6. Meirelles Cd-M, Gomes PSC. Acute effects of resistance exercise on energy expenditure:
revisiting the impact of the training variables. Rev Bras Med Esporte 2004;4:131-138.
7. Poehlman ET. A review: exercise and its influence on resting energy metabolism in man. Med Sci
Sport Exerc 1989;21:515-525.
8. Ratamess NA, Faigenbaum AD, Hoffman JR, Kang J. Self selected resistance training intensity in
healthy women: the influence of a personal trainer. J Strength Cond Res 2008;22:103-111.
9. Robergs RA, Gordon T, Reynolds J, Walker TB. Walker. Energy expenditure during the bench
press and squat exercises. J Strength Cond Res 2007;21:123-130.
10. Scott CB, Croteau A, Ravlo T. Energy expenditure before, during and after the bench press. J
Strength Cond Res 2009; 23:611-618.
11. Scott CB, Leighton B, Ahearn K, McManus J. Aerobic, anaerobic and EPOC energy expenditure
of muscular endurance and strength: 1-set of bench press to muscular fatigue (in press, J
Strength Cond Res).
12. Scott CB. A Primer for the Exercise and Nutritional Sciences: Thermodynamics,
Bioenergetics, Metabolism. Totowa, NJ: Humana Press, 2008.
13. Scott CB. Contribution of anaerobic energy expenditure to whole body thermogenesis. Nutr
Metab (online) 2005;2(14).
10
14. Tamaki T, Uchiyama S, Tamura T, Nakano S. (1994). Changes in muscle oxygenation during
weight lifting exercise. Eur J Appl Physiol. 1994;68:465-469.
15. Zoladz JA, Gladden LB, Hogan MC, Niekarz Z, Grassi B. Progressive recruitment of muscle fibers
is not necessary for the slow component of VO 2 kinetics. J Appl Physiol. 2008;105:575-580.
Disclaimer
The opinions expressed in JEPonline are those of the authors and are not attributable to JEPonline ,
the editorial staff or the ASEP organization.
11